Computer systems, information processing methods, and programs

Abstract

In gating, gates are set even in areas where the measurement data is sparse. [Solution] The computer system stores a dataset consisting of measurement data including measurement results of the intensity of multiple types of signals of particles contained in a sample measured using flow cytometry, and a model that accepts feature quantities of regions generated by dividing the observation space based on the distribution characteristics of the measurement data in the observation space, with the signal intensity as a parameter, as input, and outputs the probability of the class to which each coordinate in the observation space belongs. The computer system maps the measurement data to the observation space, divides the observation space into multiple regions based on the distribution characteristics of the measurement data in the observation space, calculates feature quantities for each region, inputs the feature quantities of each region into the model, and sets gates based on the probability of the class to which each coordinate in the observation space belongs, which is output from the model.

Description

Technical Field

[0001] The present invention relates to gating in data analysis of flow cytometry.

Background Art

[0002] Examinations of lesions using flow cytometry are increasing. In the data analysis of flow cytometry, gating that classifies particles into common groups under arbitrary conditions is important based on the measurement results. In gating, a gate that serves as the boundary of the group is set. For example, Patent Document 1 is known as a technique related to gating.

[0003] Patent Document 1 describes that "this method includes generating an image for each of the first and second sets of flow cytometer data. In some examples, generating an image includes binning the data into two-dimensional bins and assigning a shade to each bin such that the bins are represented by pixels. In some examples, the method includes maximizing the similarity to the second set of flow cytometer data by warping the generated image of the first set of flow cytometer data with a computer-implemented algorithm, and applying the same transformation to a training gate. In some embodiments, the method includes overlaying an adjusted training gate on the generated image of the second set of flow cytometer data. A system and a computer-readable medium for adjusting the training gate are also provided."

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Traditional gating methods employ clustering techniques. Clustering classifies data into groups based on similarities between data points in the feature space (observation space). However, in areas where data is not densely concentrated, the data in those areas may not be classified into any group. In applications such as lesion examinations, analyzing areas with low data density is crucial. Therefore, traditional gating methods can be difficult in such cases.

[0006] Similar challenges exist with parameter discriminant analysis and gating through threshold optimization.

[0007] The present invention aims to provide a method that enables the setting of gates even in regions where the data in the observation space is sparse. [Means for solving the problem]

[0008] A typical example of the invention disclosed in this application is as follows: a computer system comprising a processor and a memory device connected to the processor, wherein the memory device stores a dataset consisting of measurement data including measurement results of the intensity of multiple types of signals of particles contained in a sample measured using flow cytometry, and a machine learning model that receives as input feature quantities of a first region generated by dividing the observation space based on the distribution characteristics of the measurement data in the observation space, with the intensity of two or more types of signals as parameters, and outputs the probability of the class to which each coordinate in the observation space belongs, and the processor maps a plurality of the measurement data to an observation space with the intensity of a type of signal selected by the user as a parameter, divides the observation space into a plurality of first regions based on the distribution characteristics of the measurement data in the observation space, calculates feature quantities for each of the plurality of first regions, inputs each feature quantity of the plurality of first regions into the machine learning model, and sets gates based on the probability of the class to which each coordinate in the observation space belongs output from the machine learning model. [Effects of the Invention]

[0009] According to the present invention, gates can be set even in areas where measurement data is sparse. Other problems, configurations, and effects will be clarified by the following description of the embodiments. [Brief explanation of the drawing]

[0010] [Figure 1] This is a diagram showing an example of the system configuration of Example 1. [Figure 2] This figure shows an example of the functional configuration of the computer in Example 1. [Figure 3A] This figure shows an example of the measurement data database for Example 1. [Figure 3B] This figure shows an example of the measurement data database for Example 1. [Figure 4] This figure shows an example of the model database for Example 1. [Figure 5] This figure shows an example of a configuration information database for Example 1. [Figure 6] This is a flowchart illustrating an example of the analysis process performed by the computer in Example 1. [Figure 7] This figure shows an example of the partitioning of the observation space performed by the computer in Example 1. [Figure 8] This figure shows an example of the information stored by the computer in Example 1. [Figure 9] This figure shows an example of a screen displayed by the computer in Example 1. [Figure 10] This figure shows an example of the functional configuration of the computer in Example 2. [Figure 11] This is a flowchart illustrating an example of the analysis process performed by the computer in Example 2. [Modes for carrying out the invention]

[0011] The embodiments of the present invention will be described below with reference to the drawings. However, the present invention is not to be construed as being limited to the embodiments described below. It will be readily apparent to those skilled in the art that the specific configuration can be modified without departing from the spirit or intent of the present invention.

[0012] In the configuration of the invention described below, the same or similar configurations or functions are denoted by the same reference numerals, and redundant descriptions are omitted.

[0013] The notations such as "first", "second", "third", etc. in this specification and the like are attached for identifying components, and do not necessarily limit the number or order.

[0014] The positions, sizes, shapes, and ranges, etc. of each configuration shown in the drawings and the like may not represent the actual positions, sizes, shapes, and ranges, etc. in order to facilitate the understanding of the invention. Therefore, in the present invention, it is not limited to the positions, sizes, shapes, and ranges, etc. disclosed in the drawings and the like.

Example

[0015] FIG. 1 is a diagram showing a configuration example of the system of Example 1.

[0016] The system is composed of a computer 100 and a flow cytometer 101. The computer 100 and the flow cytometer 101 are connected via a network such as a LAN (Local Area Network), for example.

[0017] The flow cytometer 101 irradiates a sample containing particles such as cells with laser light, and measures the signal intensities of a plurality of types of light (measurement items) such as scattered light and fluorescence emission of the particles. The flow cytometer 101 inputs measurement data including the measurement results of each measurement item of the particles to the computer 100. That is, the same number of measurement data as the number of measured particles is input to the computer 100. Hereinafter, a plurality of measurement data is also referred to as a measurement data set.

[0018] In Example 1, the examination of cells of a patient undergoing treatment of a disease will be described as an example. In this case, a specimen such as blood becomes a sample.

[0019] Computer 100 performs analysis using the measurement dataset. Computer 100 includes a processor 110, a storage device 111, a communication device 112, an input device 113, and an output device 114. Each hardware element is connected via a bus (not shown).

[0020] The processor 110 executes a program stored in the memory device 111. By executing processing according to the program, the processor 110 operates as a functional unit (module) that realizes a specific function. In the following description, when the processing is described with a functional unit as the subject, it indicates that the processor 110 is executing a program that realizes that functional unit.

[0021] The storage device 111 is memory or the like, and stores the program executed by the processor 110 and the information used by the program. It may also include storage media such as an HDD (Hard Disk Drive) and an SSD (Solid State Drive) for storing databases, etc.

[0022] The communication device 112 communicates with external devices via a network. The input device 113 is a keyboard, mouse, touch panel, etc. The output device 114 is a display, etc.

[0023] Furthermore, the computer 100 may acquire measurement data from an external system, such as a cloud system used by the user.

[0024] Figure 2 shows an example of the functional configuration of the computer 100 in Example 1.

[0025] The computer 100 comprises an information processing unit 200, a measurement data acquisition unit 201, an input unit 202, an output unit 203, and a storage unit 204.

[0026] The measurement data acquisition unit 201 acquires measurement data sets. The input unit 202 accepts various inputs to the computer 100. The output unit 203 outputs data managed by the computer 100 and data processed by the computer 100.

[0027] The storage unit 204 manages various types of data. In Embodiment 1, the storage unit 204 manages a measurement data database for storing measurement datasets, a model database for storing machine learning models, and a configuration information database for storing analysis configuration information. Each database is stored in the storage device 111.

[0028] A machine learning model is, for example, a Transformer. The machine learning model takes as input the features of regions (occupied regions) generated by dividing the observation space based on the distribution characteristics of measurement data in the observation space, with the intensities of two or more signals as parameters, and outputs the probability of the class to which each coordinate in the observation space belongs.

[0029] By pre-generating a model that classifies the coordinates of the observation space based on the features of the occupied region, it becomes possible to set gates with high accuracy and speed.

[0030] Since the number of occupied regions is not unique, the machine learning model is assumed to be a model that can accept a sufficient number of inputs.

[0031] Furthermore, the machine learning model may accept new features calculated using the features of the occupied regions, such as the similarity of area or shape between occupied regions, as input.

[0032] The information processing unit 200 performs analysis using the measurement dataset. The information processing unit 200 includes a control unit 210, a preprocessing unit 211, a region division unit 212, a gate candidate generation unit 213, a gate setting unit 214, and an analysis unit 215.

[0033] The control unit 210 inputs various information to other functional units. The preprocessing unit 211 performs preprocessing on the measurement data. For example, data formatting and rounding of data values ​​are performed as preprocessing.

[0034] The region division unit 212 generates a scatter plot (dot plot) by plotting the measurement data in an observation space that uses the intensity of two or more types of light signals as parameters. The region division unit 212 also divides the observation space into multiple occupied regions based on the distribution of the measurement data in the observation space.

[0035] The gate candidate generation unit 213 inputs the feature quantities of each occupied region in the observation space into a machine learning model and obtains the probability of which class each coordinate in the observation space belongs to. For each class, the gate candidate generation unit 213 generates gate candidates based on the probability that the coordinate belongs to the class. The gate setting unit 214 sets the gates based on the gate candidates.

[0036] The analysis unit 215 performs analysis processing such as classification and counting of measurement data based on the configured gates.

[0037] This specification explains the observation space using the intensities of two types of light signals as parameters.

[0038] Figures 3A and 3B show an example of the measurement data database for Example 1.

[0039] The measurement data database stores tables 300 and 310.

[0040] Table 300, shown in Figure 3A, is a table for managing samples. Table 310, shown in Figure 3B, is a table for managing measurement data for samples. Each table 310 corresponds to one measurement data set. Each table 310 is assigned an identification ID for the sample.

[0041] Table 300 stores entries including ID 301, patient 302, patient characteristics 303, disease 304, and specimen type 305.

[0042] ID301 is a field that stores the identification ID of the specimen. Patient302 is a field that stores the name, etc., for identifying the patient. Patient characteristics303 is a group of fields that stores attributes representing the individual characteristics of the patient, such as age and gender. Disease304 is a field that stores the name, etc., for identifying the disease. Specimen type305 is a field that stores the type of specimen. Specimen type305 may store, for example, blood and bone marrow fluid.

[0043] Note that patient characteristics 303 may be managed in a separate table from table 300.

[0044] Table 310 stores entries that include ID 311 and measurement result 312.

[0045] ID311 is a field that stores the identification ID of the measurement data. Measurement result312 is a group of fields that store the measurement results, i.e., the measurement results for each measurement item of the particle. Measurement result312 includes fields that store signal intensities such as FSC (Forward Scatter), SSC (Side Scatter), and fluorescence at any wavelength. Note that the present invention is not limited to the type of light (measurement item) to be measured.

[0046] Figure 4 shows an example of the model database for Example 1.

[0047] The model database stores table 400. Table 400 is a table for managing machine learning models. Table 400 includes ID 401, model 402, input 403, observation space 404, and sample features 405.

[0048] ID401 is a field that stores the identification ID of the machine learning model. Model402 is a field that stores the machine learning model. Input403 is a field that stores the type of feature of the occupied region to be input to the machine learning model. Observation space404 is a field that stores the parameters (measurement items) that define the observation space. Sample features405 is a field that stores information about the characteristics of the sample. For example, the type of disease may be stored in Sample features405.

[0049] The observation space 404 and sample features 405 are fields used to select a machine learning model.

[0050] Furthermore, the entry may include a field to store information about the user performing the analysis. By managing machine learning models for each user, it is possible to support the setting of gates according to the user's purpose, characteristics, etc.

[0051] Figure 5 shows an example of a configuration information database from Example 1.

[0052] The configuration information database stores table 500. Table 500 is a table for managing the configuration information of the analysis. Table 500 stores entries including ID 501, first parameter 502, second parameter 503, and sample ID 504.

[0053] ID501 is a field that stores the identification ID of the measurement data. The first parameter 502 and the second parameter 503 are fields that store the measurement items that define the observation space. Sample ID 504 is a field that stores the identification ID of the sample.

[0054] Note that Figure 5 shows the analysis settings information without including information about the characteristics of the sample. If information about the characteristics of the sample is to be included, the entry in Table 500 will include a field to store that information.

[0055] Figure 6 is a flowchart illustrating an example of the analysis process performed by the computer 100 in Example 1. Figure 7 is a diagram illustrating an example of the division of the observation space performed by the computer 100 in Example 1. Figure 8 is a diagram illustrating an example of the information stored by the computer 100 in Example 1. Figure 9 is a diagram illustrating an example of the screen displayed by the computer 100 in Example 1.

[0056] The user performing the analysis of the measurement data inputs an analysis request, including the sample identification ID and analysis setting information, via the input unit 202 of the computer 100.

[0057] The memory unit 204 registers the input analysis setting information in the setting information database and also retrieves the measurement data set of the specified sample from the measurement data database (step S101).

[0058] When acquiring the measurement dataset, the storage unit 204 refers to table 300 and retrieves an entry with the identification ID of the sample specified as ID 301. The storage unit 204 also retrieves table 310, which contains the identification ID of the specified sample.

[0059] The memory unit 204 inputs the analysis setting information and measurement data set to the information processing unit 200. The control unit 210 of the information processing unit 200 inputs the measurement data set to the preprocessing unit 211, and also inputs the analysis setting information to the region division unit 212 and the gate candidate generation unit 213.

[0060] The preprocessing unit 211 of the information processing unit 200 performs preprocessing on the measurement dataset (step S102). The preprocessing unit 211 inputs the preprocessed measurement dataset to the region division unit 212.

[0061] The region segmentation unit 212 of the information processing unit 200 calculates the feature quantities of the occupied region based on the measurement dataset and the analysis setting information (step S103). Specifically, the following processes are performed.

[0062] (S103-1) The region division unit 212 sets the observation space based on the analysis setting information. The region division unit 212 generates a scatter plot (dot plot) by plotting the measurement data in the observation space.

[0063] (S103-2) The domain division unit 212 divides the observation space into multiple occupied regions based on the distribution characteristics of the measurement data in the observation space. For example, as shown in Figure 7, the domain division unit 212 divides the observation space into multiple occupied regions based on the Voronoi partitioning method. The measurement data becomes the source points for the Voronoi partitioning. Figure 7 shows a Voronoi diagram 710 obtained by dividing the scatter plot 700 using the Voronoi partitioning method. The Voronoi diagram 710 is the result of the partitioning.

[0064] (S103-3) The region division unit 212 calculates the feature quantities of each occupied region. For example, area, shape, circularity, etc., are calculated as feature quantities of the occupied region.

[0065] (S103-4) The region division unit 212 instructs the storage unit 204 to record the feature quantities of each occupied region. The storage unit 204 generates a table 800 with a data structure as shown in Figure 8 and adds entries for each occupied region. The storage unit 204 sets the identification ID of the measurement data in the ID 801 of each entry. The storage unit 204 sets the identification ID of the measurement data corresponding to the button included in the occupied region in the data ID 802 of each entry, and sets the feature quantity of the occupied region corresponding to the measurement data in the region feature quantity 804 of each entry.

[0066] (S103-5) The region division unit 212 outputs a scatter plot, a Voronoi diagram (division result), and a table 800 to the gate candidate generation unit 213.

[0067] The above is a description of the process in step S103. Note that methods other than Voronoi partitioning may be used to partition the observation space. For example, a partitioning method that utilizes the maximum likelihood estimation results of an adaptive kernel according to the shape of the distribution can be considered.

[0068] The gate candidate generation unit 213 of the information processing unit 200 generates gate candidates using the feature quantities of the occupied region (step S104). Specifically, the following processing is performed.

[0069] (S104-1) The gate candidate generation unit 213 obtains information on the parameters that define the observation space included in the analysis setting information. The gate candidate generation unit 213 refers to the table 400 stored in the model database and searches for an entry in the observation space 404 where the obtained parameters are set. If the analysis setting information includes characteristics of the sample, etc., these items are also taken into consideration.

[0070] (S104-2) The gate candidate generation unit 213 obtains a machine learning model from the model 402 of the searched entry.

[0071] If multiple entries exist, the gate candidate generation unit 213 may present selectable machine learning models via the output unit 203, and the user may choose one. Alternatively, the gate candidate generation unit 213 may make a random selection or a selection based on predetermined selection rules.

[0072] (S104-3) The gate candidate generation unit 213 inputs the features of the occupied region into the machine learning model according to the type of features of the occupied region set in the input 403 of the searched entry.

[0073] (S104-4) The gate candidate generation unit 213 identifies classes based on the output of the machine learning model and selects one class from among the identified classes.

[0074] (S104-5) The gate candidate generation unit 213 sets a threshold. In this embodiment, it is assumed that a list of thresholds has been registered in advance. The gate candidate generation unit 213 selects one threshold from the list.

[0075] (S104-6) The gate candidate generation unit 213 generates gate candidates for the selected class, based on the output of the machine learning model, the boundaries of regions formed by coordinate points whose probability is above a threshold. Specifically, the gate candidate generation unit 213 records data that associates the coordinates of the boundary to be a gate candidate with the threshold and the class.

[0076] Furthermore, if there are multiple regions formed by coordinate points whose probability is above a threshold, multiple gate candidates will be generated.

[0077] (S104-7) The gate candidate generation unit 213 determines whether processing has been completed for all thresholds registered in the list. If processing has not been completed for all thresholds registered in the list, the gate candidate generation unit 213 returns to S104-5. If processing has been completed for all thresholds registered in the list, the gate candidate generation unit 213 proceeds to S104-8.

[0078] (S104-8) The gate candidate generation unit 213 determines whether processing has been completed for all classes. If processing has not been completed for all classes, the gate candidate generation unit 213 returns to S104-4. If processing has been completed for all classes, the gate candidate generation unit 213 proceeds to S104-9.

[0079] (S104-9) The gate candidate generation unit 213 outputs a scatter plot and gate candidates for each class to the gate setting unit 214.

[0080] The above is a description of the process in step S104. Note that the class loop and the threshold loop are interchangeable.

[0081] The gate setting unit 214 sets the gate based on the gate candidates (step S105). Specifically, the following processes are performed.

[0082] (S105-1) The gate setting unit 214 displays the screen 900 via the output unit 203.

[0083] Now, let's explain screen 900. Screen 900 includes selection fields 901 and 902, display field 903, and buttons 904 and 905.

[0084] Selection field 901 is for selecting a class. The selectable classes are displayed in selection field 901. In addition, a symbol indicating that a class has been processed is displayed for classes for which a gate has been set. Selection field 902 is for selecting gate candidates for the class selected in selection field 901. Display field 903 is a field that displays a magnified view of the gate candidates selected in selection field 902.

[0085] (S105-2) When the user selects a class from the selection field 901, the gate setting unit 214 displays gate candidates for the selected class in the selection field 902. In Figure 9, the elliptical boundaries represent the gate candidates.

[0086] (S105-3) When the user selects a gate candidate from the selection field 902, the gate setting unit 214 displays the selected gate candidate in an enlarged view in the display field 903. If there is a class for which gates are already set, the gate setting unit 214 may overlay the gates of that class in the display field 903.

[0087] To modify a gate candidate, the user operates button 904 to modify the gate candidate displayed in display area 903. To set a gate candidate displayed in display area 903 as a gate, the user operates button 905.

[0088] (S105-4) When the gate settings are complete for all classes, the gate setting unit 214 displays a screen via the output unit 203 to confirm whether the gate settings are complete or not. If the gate settings are complete, the gate setting unit 214 outputs the scatter plot and gates to the analysis unit 215.

[0089] The above is a description of the process in step S105.

[0090] The analysis unit 215 counts the number of measurement data points in the region formed by the set gate (step S106) and outputs the count result via the output unit 203 (step S107).

[0091] Note that only one threshold can be set. In this case, the gate candidate will be set as the gate. The analysis unit 215 may also perform analyses to estimate the remaining state of the lesion based on the number of measurement data in a predetermined area.

[0092] As explained above, computer 100 generates gate candidates using feature quantities of occupied regions obtained by dividing the observation space based on the characteristics of the distribution of measurement data in the observation space. Unlike conventional clustering methods, this method does not focus solely on the density of measurement data, so it can generate gate candidates even in spaces where the measurement data is sparse. Furthermore, the user can quickly and easily set the desired gate by modifying the presented gate candidates as needed and selecting one gate candidate from among the presented candidates. [Examples]

[0093] In Example 2, the computer 100 differs from Example 1 in that it uses the gate setting results to train the machine learning model. Below, Example 2 will be described focusing on the differences from Example 1.

[0094] The system configuration of Example 2 is the same as that of Example 1. The hardware configuration of computer 100 in Example 2 is the same as that of Example 1.

[0095] In Example 2, the functional configuration of the computer 100 differs from that of Example 1. Figure 10 shows an example of the functional configuration of the computer 100 in Example 2. The computer 100 in Example 2 newly includes a learning unit 205. In addition, the storage unit 204 in Example 2 differs from that in Example 1 in that it manages a gate database that stores the gate setting results.

[0096] Figure 11 is a flowchart illustrating an example of the analysis process performed by the computer 100 in Example 2.

[0097] After the processing in step S105, the gate setting unit 214 determines whether or not a gate candidate modified by the user for at least one class has been set as a gate (step S151).

[0098] If the gate candidate modified by the user is not set as a gate, the computer 100 proceeds to step S106.

[0099] If a gate candidate modified by the user is set as a gate for at least one class, the gate setting unit 214 records the model identification ID, table 800, and data associating the gate in the gate database (step S152). The computer 100 then proceeds to step S106.

[0100] Alternatively, computer 100 may skip the processing in step S151, execute the processing in step S152, and then proceed to step S106.

[0101] The learning unit 205 performs learning processing when data is recorded in the gate database, or periodically. The learning unit 205 uses the data stored in the gate database to train a machine learning model. The learning method is not limited.

[0102] According to Example 2, the accuracy of a machine learning model can be improved through continuous learning.

[0103] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. Furthermore, for example, the embodiments described above are detailed explanations of the configuration in order to clearly illustrate the present invention, and are not necessarily limited to those having all the configurations described. In addition, some of the configurations in each embodiment can be added to, deleted from, or replaced with other configurations.

[0104] Furthermore, each of the above-mentioned configurations, functions, processing units, processing means, etc., may be implemented in hardware, in whole or in part, for example, by designing them as integrated circuits. The present invention can also be implemented by software program code that realizes the functions of the embodiment. In this case, a storage medium on which the program code is recorded is provided to a computer, and the processor of that computer reads the program code stored in the storage medium. In this case, the program code read from the storage medium itself realizes the functions of the embodiment described above, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media used to supply such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs (Solid State Drives), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, and the like.

[0105] Furthermore, the program code that implements the functions described in this embodiment can be implemented in a wide range of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, Python, and Java (registered trademark).

[0106] Furthermore, the program code for the software that implements the functions of the embodiment may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the computer's processor may read and execute the program code stored in the storage means or storage medium.

[0107] In the above-described embodiment, the control lines and information lines shown are those deemed necessary for explanation and do not necessarily represent all control lines and information lines in the actual product. All components may be interconnected. [Explanation of Symbols]

[0108] 100 calculator 101 Flow cytometer 110 processors 111 Storage device 112 Communication equipment 113 Input device 114 Output device 200 Information Processing Unit 201 Measurement Data Acquisition Unit 202 Input Section 203 Output section 204 Storage section 205 Learning Department 210 Control Unit 211 Pre-processing section 212 Area division part 213 Gate Candidate Generation Unit 214 Gate setting section 215 Analysis Department 300, 310, 400, 500, 800 tables 700 scatter plot 710 Division result 900 screens

Claims

1. A computer system comprising a processor and a storage device connected to the processor, The aforementioned storage device is A dataset consisting of measurement data including the results of measuring the intensity of multiple types of signals from particles contained in a sample, measured using flow cytometry, A machine learning model is stored which accepts as input the feature quantities of a first region generated by dividing the observation space based on the distribution characteristics of the measurement data in the observation space, with the intensities of two or more types of signals as parameters, and outputs the probability of the class to which each coordinate in the observation space belongs. The aforementioned processor, Multiple measurement data are mapped to an observation space whose parameters are the intensity of the type of signal selected by the user. Based on the characteristics of the distribution of the measurement data in the observation space, the observation space is divided into a plurality of first regions. The feature quantities of each of the aforementioned plurality of first regions are calculated, Each of the feature quantities of the plurality of first domains is input to the machine learning model. A computer system characterized by setting gates based on the probability of the class to which each coordinate in the observation space, output from the machine learning model, belongs.

2. A computer system according to claim 1, The aforementioned processor, The process of setting a threshold and generating gate candidates for the boundary of a second region formed by the coordinates whose probability is greater than the threshold, based on the probability of the class to which each coordinate in the observation space output from the machine learning model belongs, is repeatedly performed. For each of the aforementioned classes, an interface is presented to the user to allow them to select one of the gate candidates from among the multiple gate candidates of the aforementioned class as the gate. A computer system characterized by setting the gate for each class based on the input from the user received via the interface.

3. A computer system according to claim 2, The computer system is characterized in that the processor sets the user-modified gate candidate received via the interface as the gate.

4. A computer system according to claim 3, The aforementioned processor, The data relating the feature quantities of the first region to the gate candidates modified by the user is stored in the storage device. A computer system characterized by performing the training process of the machine learning model using the aforementioned data.

5. A computer system according to claim 1, The processor is a computing system characterized by dividing the observation space into a plurality of first regions based on the Voronoi partitioning method.

6. A computer system according to claim 1, The aforementioned processor, The number of measurement data included in the third region formed by the gate is counted. A computer system characterized by outputting the result of the aforementioned count.

7. A computer system according to claim 1, Storing multiple machine learning models, The machine learning model is managed in association with the type of parameters defining the observation space and the type of sample. The computer system is characterized in that the processor selects the machine learning model to be used based on the type of parameters defining the observation space and the type of sample.

8. A method of information processing performed by a computer system, The aforementioned computer system, It has a processor and a storage device connected to the processor, The aforementioned storage device is A dataset consisting of measurement data including the intensity of multiple types of signals for each of multiple particles contained in a sample measured using flow cytometry, A machine learning model is stored which accepts as input the feature quantities of a first region generated by dividing the observation space based on the distribution characteristics of the measurement data in the observation space, with the intensities of two or more types of signals as parameters, and outputs the probability of the class to which each coordinate in the observation space belongs. The aforementioned information processing method is The processor performs the steps of mapping a plurality of the measurement data into an observation space whose parameters are the intensity of a signal of a type selected by the user, The processor divides the observation space into a plurality of first regions based on the characteristics of the distribution of the measurement data in the observation space. The processor performs the step of calculating the feature quantities of each of the plurality of first regions, The processor inputs the feature quantities of each of the plurality of first regions into the machine learning model. An information processing method characterized in that the processor sets a gate based on the probability of the class to which each coordinate in the observation space output from the machine learning model belongs.

9. A program to be executed by a computer, The aforementioned computer is A dataset consisting of measurement data including the intensity of multiple types of signals for each of multiple particles contained in the sample, measured using flow cytometry, A machine learning model is maintained which accepts a feature quantity of a first region, generated by dividing the observation space based on the distribution characteristics of the measurement data in the observation space, with the intensities of two or more types of signals as parameters, as input, and outputs the probability of the class to which each coordinate in the observation space belongs. The aforementioned program, A procedure for mapping multiple measurement data into an observation space whose parameters are the intensity of a signal of a type selected by the user, A procedure for dividing the observation space into a plurality of first regions based on the characteristics of the distribution of the measurement data in the observation space, A procedure for calculating the feature quantities of each of the aforementioned multiple first regions, A procedure for inputting the feature quantities of each of the multiple first domains into the machine learning model, A program characterized by causing the computer to perform a procedure for setting gates based on the probability of the class to which each coordinate in the observation space, output from the machine learning model, belongs.

Images